Figure 2 - uploaded by S. Kevin Zhou
Content may be subject to copyright.
The LogitBoost algorithm [4]. 

The LogitBoost algorithm [4]. 

Source publication
Conference Paper
Full-text available
We tackle the problem of automatically classifying cardiac view for an echocardiographic sequence as a multiclass object detection. As a solution, we present an imagebased multiclass boosting procedure. In contrast with conventional approaches for multiple object detection that train multiple binary classifiers, one per object, we learn only one mu...

Contexts in source publication

Context 1
... Figure 4 illustrate fitting the piecewise constant function to a set of unweighted data points. In the step ( ∆ ) in Figure 2, the most relevant feature with the smallest error is singled out to maximally reduce the classification error. In the literature, simple binary regression stumps are used [17, 15]. However, it is easy to show that the piecewise constant function combines multiple binary regression stumps, therefore enhancing the modeling capability of the weak learners and consequently improving training speed. In this section, we address two useful extensions of the multiclass boosting algorithm. In the first extension, we show how to learn a tree structure which offers efficiency in both training and testing. In the second extension, we propose to train a cascade structure to deal with the vexing background class that has voluminous examples. Empirical evidence tells that often, in the midst of boosting a multiclass classifier, one class (or several classes) has been completely separated from the remaining ones and further boosting yields no additional improvement in terms of the classification accuracy. This fact can be utilized for efficient training. To this end, we propose to train a tree structure. Figure 5 gives a simple example to illustrate the tree training. Suppose we are given a 3-class problem ( C 1 , C 2 , and C 3 ). After several boosting iterations, we find that the class C 1 has been classified correctly. We stop training and store the output functions as { F 1 ,j ( x ); j = 1 , 2 , 3 } , which forms the first layer of the tree that merges the classes C 2 and C 3 . Next, for the second layer of the tree, we continue to train a binary classifier that separates C 2 and C 3 and store the output functions as { F 2 ,j ( x ); j = 2 , 3 } . To calculate the posterior probability of the class label, we first computer the posterior probability for each tree layer. For example, for the first layer, we ...
Context 2
... adopt the multiclass version of the influential boosting algorithm proposed by Friedman et al . [4], the so-called LogitBoost algorithm. The output of the LogitBoost algorithm is a set of ( J + 1) learned response functions { F j ( x ); j = 0 , 1 , . . . , J } ; each F j ( x ) is a linear combi- nation of the set of weak learners, thereby implying that the F j ( x ) functions automatically share features [15]. Figure 2 illustrates the LogitBoost algorithm, which fits an additive symmetric logistic model to achieve maximum likeli- hood using adaptive quasi-Newton steps, whereas the Ad- aBoost.MH algorithm essentially uses the one against the rest rule by fitting ( J + 1) uncoupled additive logistic models. See [4] for detailed justification. The final classification result is determined ...
Context 3
... advantage of using the LogitBoost algorithm is that it naturally provides a way to calculate the posterior distri- bution of class label: (4) where F j ′ ( x ) = F j ( x ) − F 0 ( x ) . The key of the LogitBoost algorithm is the step ( ∆ ) in Figure 2. Mathematically, it solves the following optimiza- tion ...

Similar publications

Preprint
Full-text available
Convolution Neural Networks (CNN) have been extremely successful in solving intensive computer vision tasks. The convolutional filters used in CNNs have played a major role in this success, by extracting useful features from the inputs. Recently researchers have tried to boost the performance of CNNs by re-calibrating the feature maps produced by t...
Preprint
Full-text available
Convolution Neural Networks (CNN) have been extremely successful in solving intensive computer vision tasks. The convolutional filters used in CNNs have played a major role in this success, by extracting useful features from the inputs. Recently researchers have tried to boost the performance of CNNs by re-calibrating the feature maps produced by t...
Article
Full-text available
The Adaboost (Freund and Schapire, Eur. Conf. Comput. Learn. Theory 23–37, 1995) chooses a good set of weak classifiers in rounds. On each round, it chooses the optimal classifier (optimal feature and its threshold value) by minimizing the weighted error of classification. It also reweights training data so that the next round would focus on data t...
Conference Paper
Full-text available
Viola et al. have introduced a fast object detection scheme based on a boosted cascade of haar-like features. In this paper, we introduce a novel ternary feature that enriches the diversity and the flexibility significantly over haar-like features. We also introduce a new genetic algorithm (GA) based method for training effective ternary features t...

Citations

... Some prior works had proposed feature extraction techniques before applying machine learning algorithms. For instance, Zhou et al. [46] considered the LogitBoost algorithm to automatically classify cardiac view as a multiclass object detection. Similarly, Park et al. [47] presented an automatic system for cardiac view classification by characterizing local and global evidence, specific knowledge, and applying Logit-Boost. ...
Article
Full-text available
Detecting cardiac abnormalities between 14 and 28 weeks of gestation with an apical four-chamber view is a difficult undertaking. Several unfavorable factors can prevent such detection, such as the fetal heart’s relatively small size, unclear appearances in anatomical structures (e.g., shadows), and incomplete tissue boundaries. Cardiac defects without segmentation are not always straightforward to detect, so using only segmentation cannot produce defect interpretation. This paper proposes an improved semantic segmentation approach that uses a region proposal network for septal defect detection and combines two processes: contour segmentation with U-Net architecture and defect detection with Faster-RCNN architecture. The model is trained using 764 ultrasound images that include three abnormal conditions (i.e., atrial septal defect, ventricular septal defect, and atrioventricular septal defect) and normal conditions from an apical four-chamber view. The proposed model produces a satisfactory mean intersection over union, mean average precision, and dice similarity component metrics of about 75%, 87.80%, and 96.37%, respectively. Furthermore, the proposed model has also been validated on 71 unseen images in normal conditions and produces 100% sensitivity, which means that all normal conditions without septal defects can be detected effectively. The developed model has the potential to identify the fetal heart in normal and pathological settings accurately. The developed deep learning model's practical use in identifying congenital heart disorders has substantial future promise.
... Support vector machines (SVM) and linear discriminant analysis (LDA) have been used as one of the primary tools for classification by learning the decision boundaries and classifying the different views in space [36][37][38][39][40][41]. Multi-class logit-boost classifiers are also proposed for classification of the view in echocardiographic images [42,43]. Khamis et al. [44] proposed a multi-stage classification algorithm for employing spatio-temporal feature extraction and supervised dictionary learning to classify longitudinal scans namely: apical two-chamber (A2C), apical four-chamber (A4C) and apical long-axis (ALX), as shown in Figure 2. The inherent noise makes the classification challenging. ...
Article
Full-text available
Cardiovascular diseases (CVDs) remain the leading cause of death worldwide. An effective management and treatment of CVDs highly relies on accurate diagnosis of the disease. As the most common imaging technique for clinical diagnosis of the CVDs, US imaging has been intensively explored. Especially with the introduction of deep learning (DL) techniques, US imaging has advanced tremendously in recent years. Photoacoustic imaging (PAI) is one of the most promising new imaging methods in addition to the existing clinical imaging methods. It can characterize different tissue compositions based on optical absorption contrast and thus can assess the functionality of the tissue. This paper reviews some major technological developments in both US (combined with deep learning techniques) and PA imaging in the application of diagnosis of CVDs.
... It is used in treatment decisionmaking, curative effect evaluation, and genetic diseases of heart diseases. Screening and epidemiological investigations also play an important role, so it is widely used in clinical diagnosis at all levels and has become the preferred method of cardiac structure and function evaluation [1][2][3][4][5]. ...
Article
Full-text available
Cardiovascular disease remains a substantial cause of morbidity and mortality in the developed world and is becoming an increasingly important cause of death in developing countries too. While current cardiovascular treatments can assist to reduce the risk of this disease, a large number of patients still retain a high risk of experiencing a life-threatening cardiovascular event. Thus, the advent of new treatments methods capable of reducing this residual risk remains an important healthcare objective. This paper proposes a deep learning-based method for section recognition of cardiac ultrasound images of critically ill cardiac patients. A convolution neural network (CNN) is used to classify the standard ultrasound video data. The ultrasound video data is parsed into a static image, and InceptionV3 and ResNet50 networks are used to classify eight ultrasound static sections, and the ResNet50 with better classification accuracy is selected as the standard network for classification. The correlation between the ultrasound video data frames is used to construct the ResNet50 + LSTM model. Next, the time-series features of the two-dimensional image sequence are extracted and the classification of the ultrasound section video data is realized. Experimental results show that the proposed cardiac ultrasound image recognition model has good performance and can meet the requirements of clinical section classification accuracy.
... 2,3 Application of machine learning algorithms in computer vision has improved the accuracy and time-efficiency of automated image analysis, particularly automated interpretation of medical images. [4][5][6][7] However, traditional machine learning methods are constructed using complex processes and tend to have a restricted scope and effectiveness. 8,9 Recent advances in the design and application of deep neural networks have resulted in increased possibilities when automating medical image-based diagnosis. ...
Article
Purpose: Echocardiography is the most commonly used modality for assessing the heart in clinical practice. In an echocardiographic exam, an ultrasound probe samples the heart from different orientations and positions, thereby creating different viewpoints for assessing the cardiac function. The determination of the probe viewpoint forms an essential step in automatic echocardiographic image analysis. Approach: In this study, convolutional neural networks are used for the automated identification of 14 different anatomical echocardiographic views (larger than any previous study) in a dataset of 8732 videos acquired from 374 patients. Differentiable architecture search approach was utilized to design small neural network architectures for rapid inference while maintaining high accuracy. The impact of the image quality and resolution, size of the training dataset, and number of echocardiographic view classes on the efficacy of the models were also investigated. Results: In contrast to the deeper classification architectures, the proposed models had significantly lower number of trainable parameters (up to 99.9% reduction), achieved comparable classification performance (accuracy 88.4% to 96%, precision 87.8% to 95.2%, recall 87.1% to 95.1%) and real-time performance with inference time per image of 3.6 to 12.6 ms. Conclusion: Compared with the standard classification neural network architectures, the proposed models are faster and achieve comparable classification performance. They also require less training data. Such models can be used for real-time detection of the standard views.
... There are several prior works studying cardiac motion analysis using ultrasound images. Variable image features obtained from 2D cardiac ultrasound images have been used for myocardial infarction (MI) estimation [3]. Carneiro et al. [4], [5] used rectangular filters to train detectors for different anatomical structures recognition and the probabilistic boosting tree cascade architecture for object detection. ...
... The main strength of this method is that it does not require constructing spatial and temporal models for each cardiac view. Other conventional-based methods for view classification are presented in [38] (bag of visual words with SVM), [39] (gradient features and logistic trees), [40] (visual features and boosting), [41] (B-spline and thresholding), and [42] (histogram features and neural network). ...
Article
Echocardiography (echo) is a critical tool in diagnosing various cardiovascular diseases. Despite its diagnostic and prognostic value, interpretation and analysis of echo images are still widely performed manually by echocardiographers. A plethora of algorithms has been proposed to analyze medical ultrasound data using signal processing and machine learning techniques. These algorithms provided opportunities for developing automated echo analysis and interpretation systems. The automated approach can significantly assist in decreasing the variability and burden associated with manual image measurements. In this paper, we review the state-of-the-art automatic methods for analyzing echocardiography data. Particularly, we comprehensively and systematically review existing methods of four major tasks: echo quality assessment, view classification, boundary segmentation, and disease diagnosis. Our review covers three echo imaging modes, which are B-mode, M-mode, and Doppler. We also discuss the challenges and limitations of current methods and outline the most pressing directions for future research. In summary, this review presents the current status of automatic echo analysis and discusses the challenges that need to be addressed to obtain robust systems suitable for efficient use in clinical settings or point-of-care testing.
... To the best of our knowledge, this is the first work to attempt to automate analysis of fetal cardiac ultrasound videos. Previous authors have successfully performed view detection in images obtained from adult echocardiographic images using a variety of techniques ( Agarwal et al. (2013) ; Wu et al. (2013) ; Zhou et al. (2006) ; Park et al. (2007) ; Qian et al. (2013) ; Kumar et al. (2009) ;Ebadollahi et al. (2004) ), while others have had success in automatic recognition of other fetal structures in images ( Carneiro et al. (2008) ; Rahmatullah et al. (2012) ; Namburete et al. (2013) ; Yaqub et al. (2012) ) and, more recently, in videos ( Maraci et al. (2014) ; Chen et al. (2015) ). Finally, some work has attempted to estimate a more detailed description of the adult heart in echocardiographic data in the form of boundaries ( Nascimento and Marques (2008) ; Carneiro and Nascimento (2013) ; Yang et al. (2008) ). ...
... Wu et al. (2013) employ a similar method, using 'GIST' descriptors ( Oliva and Torralba (2001) ) in 16 image blocks instead of HOG descriptors. Zhou et al. (2006) use a multi-class classifier based on LogitBoost and rectangular filters ('Haar-like' filters) in order to distinguish between apical twochamber and four-chamber views. Such global methods are not well-suited to fetal echocardiography because they assume a relatively consistent layout of frames, but in fetal imagery the position and orientation of the heart is unknown. ...
... Also, in our application, only small areas of the fetal images are relevant to view classification, and the rest of the image is taken up by the fetal abdomen and the womb. This is overcome, to some extent, in the work of Park et al. (2007) , which builds on the work in Zhou et al. (2006) by adding a left ventricle detection stage, which is then used to position the multi-class view classifier in the image. However, this relies upon the appearance of the left ventricle being fairly consistent between views, and there is unfortunately no such guarantee of consistency in the fetal views of interest to us. ...
Article
Full-text available
Interpretation of ultrasound videos of the fetal heart is crucial for the antenatal diagnosis of congenital heart disease (CHD). We believe that automated image analysis techniques could make an important contribution towards improving CHD detection rates. However, to our knowledge, no previous work has been done in this area. With this goal in mind, this paper presents a framework for tracking the key variables that describe the content of each frame of freehand 2D ultrasound scanning videos of the healthy fetal heart. This represents an important first step towards developing tools that can assist with CHD detection in abnormal cases. We argue that it is natural to approach this as a sequential Bayesian filtering problem, due to the strong prior model we have of the underlying anatomy, and the ambiguity of the appearance of structures in ultrasound images. We train classification and regression forests to predict the visibility, location and orientation of the fetal heart in the image, and the viewing plane label from each frame. We also develop a novel adaptation of regression forests for circular variables to deal with the prediction of cardiac phase. Using a particle-filtering-based method to combine predictions from multiple video frames, we demonstrate how to filter this information to give a temporally consistent output at real-time speeds. We present results on a challenging dataset gathered in a real-world clinical setting and compare to expert annotations, achieving similar levels of accuracy to the levels of inter- and intra-observer variation.
... One can learn that average recognition accuracies using the proposed two-stage algorithm are higher than other reported methods (95% (proposed method), 68.33% (Qian et al., 2013), 67.8% (Ebadollahi et al., 2004)). Also A2C and A4C recognition accuracies are higher than other reported methods (94% (proposed method), 85.75% (Aschkenasy et al., 2006), 93.6% (Park et al., 2007), 90.4% (Kevin Zhou et al., 2006), 82.5% (Balaji et al., 2015a), 88.75% (Balaji et al., 2015b) and ...
Article
As part of striving towards fully automatic cardiac functional assessment of echocardiograms, automatic classification of their standard views is essential as a pre-processing stage. The similarity among three of the routinely acquired longitudinal scans: apical two-chamber (A2C), apical four-chamber (A4C) and apical long-axis (ALX), and the noise commonly inherent to these scans - make the classification a challenge. Here we introduce a multi-stage classification algorithm that employs spatio-temporal feature extraction (Cuboid Detector) and supervised dictionary learning (LC-KSVD) approaches to uniquely enhance the automatic recognition and classification accuracy of echocardiograms. The algorithm incorporates both discrimination and labelling information to allow a discriminative and sparse representation of each view. The advantage of the spatio-temporal feature extraction as compared to spatial processing is then validated.
... For example, Kumar et al. [1] are devoted to expressing the main anatomical structure of space motion. Zhou et al. [2] and Beymer et al. [3] mention echocardiogram classification based on multiple object space relations. Then Shalbaf and Sera et al. [4,5] add the movement information on this basis, which extract and detect the feature by tracking the movements and patterns of the contour shape of the heart. ...
Article
Full-text available
In echo-cardiac clinical computer-aided diagnosis, an important step is to automatically classify echocardiography videos from different angles and different regions. We propose a kind of echocardiography video classification algorithm based on the dense trajectory and difference histograms of oriented gradients (DHOG). First, we use the dense grid method to describe feature characteristics in each frame of echocardiography sequence and then track these feature points by applying the dense optical flow. In order to overcome the influence of the rapid and irregular movement of echocardiography videos and get more robust tracking results, we also design a trajectory description algorithm which uses the derivative of the optical flow to obtain the motion trajectory information and associates the different characteristics (e.g., the trajectory shape, DHOG, HOF, and MBH) with embedded structural information of the spatiotemporal pyramid. To avoid “dimension disaster,” we apply Fisher’s vector to reduce the dimension of feature description followed by the SVM linear classifier to improve the final classification result. The average accuracy of echocardiography video classification is 77.12% for all eight viewpoints and 100% for three primary viewpoints.
... Zhou et al. [34] presented an approach based on multiple object detection. They manually defined templates based on the left ventricle (LV) orientation and size, which were used to align the data and reduce appearance variation. ...
... 2) Discussion: Most of the methods presented have peculiarities which can create constraints or extra costs in their use. For instance, some approaches [9], [19], [34] only deal with the end diastolic (ED) frame, which could limit their use in the real-time scenario (heart view shown during the examination). Waiting for the ED frame to be displayed may delay the system response. ...
... Many methods [1], [11], [18], [25], [34] also ap-ply pre-processing steps to normalize images/videos. Contrast/brightness normalization, noise reduction, alignment, and so on, usually introduce extra costs. ...
Article
Full-text available
In this paper, we explore mid-level image representations for real-time heart view plane classification of 2D echocardiogram ultrasound images. The proposed representations rely on bags of visual words, successfully used by the computer vision community in visual recognition problems. An important element of the proposed representations is the image sampling with large regions, drastically reducing the execution time of the image characterization procedure. Throughout an extensive set of experiments, we evaluate the proposed approach against different image descriptors for classifying four heart view planes. The results show that our approach is effective and efficient for the target problem, making it suitable for use in real-time setups. The proposed representations are also robust to different image transformations, e.g., downsampling, noise filtering, and different machine learning classifiers, keeping classification accuracy above 90%. Feature extraction can be performed in 30fps or 60fps in some cases. This paper also includes an in-depth review of the literature in the area of automatic echocardiogram view classification giving the reader a through comprehension of this field of study.